Sentiment Analysis of Visitor Reviews on Hotel in West Sumatera

Hotel XYZ is one of the hotels in Padang City and the only five-star hotel. Now, visitors can make hotel reservation online easily from online travel sites, one of them is from the TripAdvisor site. If was viewed from the site, it can be seen that hotel has low rating fairly if it compared with the other type of hotel. So, hotel management needs to arrange the strategy to increase the hotel rating again. This is the basic reason so that it needs to evaluate how the actual hotel market on this site, whether it is good or bad using sentiment analysis from visitor reviews data by classifying positive and negative sentiments. Classification method that used in this study is decision tree and it will be compared with other method, namely naïve bayes using k-fold cross validation. Calculation results of accuracy and precision value using decision tree give greater results than naive bayes while both methods have same recall value which indicate reliability in predicting positive data. Based on the value, it can be determined that classification results of positive sentiment are greater than negative sentiment. Thus, it can be concluded that Hotel XYZ on the TripAdvisor site get a lot of positive sentiment from visitors.